Protein structure matching by genetic algorithm

With the increase in the availability of protein structure information, the ability to provide accurate matching between two proteins or subsequences of proteins becomes more important This paper presents a solution to the protein matching problem by providing a dynamic superposition of the three-dimensional coordinate string of one protein subsequence, called the query sequence, onto the three-dimensional coordinate string of a second fixed protein sequence, using a genetic algorithm. The representation of a protein is by the ordered listing of its three-dimensional coordinates that correspond to the alpha carbon backbone structure. The solution provided by the genetic algorithm is an isometry, a mathematical transformation that moves one three-dimensional structure (the query protein string) onto another three-dimensional structure (the fixed reference protein string;). It provides a contiguous map of one protein string to another. The generic algorithm was applied to 31 proteins from the Brookhaven Protein Data Bank and found an approximate solution quickly.